A federated bridge built mainly to feed assets into Hyperliquid, not a Monad-focused product — and its 450 MON minimum deposit has a genuinely unforgiving failure mode.
Unit (also called HyperUnit) is not a Monad-native or Monad-focused bridge — it's an asset-tokenization layer built primarily to feed spot assets into Hyperliquid. It uses a lock-and-mint model run by a federated "Guardian Network" — currently a small set of named organizations using MPC/threshold key-sharing so no single party ever holds a full signing key. Guardians watch the source chain for a deposit and, on reaching consensus, mint a wrapped representation on Hyperliquid; withdrawals reverse the process. Monad is one of roughly a dozen supported source chains — alongside Bitcoin, Ethereum, Solana, Plasma, Zcash and others — a spoke, not the hub.
Each Guardian runs four modules: a Chain Services component that monitors source chains and confirms finality, a Flow Manager that enforces correct operation sequencing, a Consensus Service that enforces the quorum rule, and a Wallet Manager that executes MPC threshold signing. Private keys never exist in complete form — they're generated and held as split shares inside secure hardware enclaves, and a full key is never reconstructed even during signing. The quorum itself, though, is currently just three named entities requiring 2-of-3 agreement: Unit's own core team, Hyperliquid, and a third-party infrastructure firm called Infinite Field — a materially different trust model from "many independent Guardians," since two of the three parties are directly affiliated with the protocol that benefits from the bridge. Unit's own team has stated an intent to expand and further decentralize the Guardian set over time, which is itself an implicit acknowledgment that the current setup is a compromise, not the end state.
Unit charges zero platform fee on deposits or withdrawals — every dollar a user pays is just the two native gas costs (the source-chain transaction and the Guardian-executed destination-chain transaction), with no spread or bridge toll layered on top. The protocol also bakes compliance controls directly into the Guardian layer rather than leaving it to Hyperliquid's front end: OFAC transaction screening, IP-based geoblocking, VPN detection, rate-limiting and a collective circuit breaker for abnormal activity.
Against NEAR Intents — the other cross-chain option in this set — the two sit on opposite ends of the custody spectrum. Unit is federated custody-in-transit: between a deposit and the Guardian-minted asset appearing on Hyperliquid, funds sit locked in Guardian-controlled MPC wallets, and a 2-of-3 signature from three named parties is both necessary and sufficient to move them — the same trust family as a classic multisig bridge, implemented with MPC instead of a plain multisig contract. NEAR Intents has no pooled custody at all: each intent is independently collateralized by whichever solver fills it, so a single compromised key can't drain a shared vault the way it theoretically could with Unit's Guardian quorum. The practical tradeoff is that Unit's model buys Hyperliquid deep, purpose-built integration — native BTC and ETH tradeable directly on HyperCore — at the cost of trusting three parties with MPC key shares, while NEAR Intents is more trust-minimized per transaction but depends on a solver actually showing up with competitive liquidity for a given pair.
Deposits from Monad only accept native MON, arrive on Hyperliquid within about a minute, and — this is the important part — carry a hard minimum of 450 MON per deposit. Anything sent below that threshold results in permanent loss of funds, with no described refund mechanism.
Despite MPC key-sharing distributing signing across the Guardian set, that set is currently just three named organizations — a meaningfully weaker trust assumption than a large, permissionless validator set, and a compromise or collusion among two of the three would be sufficient to move funds. There's also a documented history of at least one operational incident, in April 2025, where a single Guardian going offline delayed Bitcoin deposits and withdrawals platform-wide — showing the system isn't yet fault-tolerant to routine node outages when the quorum is this small. And the 450 MON minimum-deposit rule is a concrete, protocol-specific way to lose funds through ordinary user error, not a generic "smart contract risk" — worth knowing before sending anything through this route from Monad.
Note: Monad is peripheral to Unit's actual design center — Hyperliquid — so Monad-side liquidity and attention may not track the health of the core product.
No — Hyperliquid's own support documentation states deposits under the 450 MON minimum on the Monad network are permanently lost, with no described recovery path. This is an unforgiving, protocol-specific floor, not a "gets refunded eventually" situation.
A 2-of-3 MPC quorum made up of three named parties: Unit's own team, Hyperliquid, and a third-party infrastructure firm called Infinite Field. No single party can move funds alone, but two of the three colluding or being compromised together is sufficient, and two of the three are directly affiliated with the destination protocol.
No — Unit takes zero platform revenue on deposits or withdrawals. You only pay the two native gas costs: the source-chain transaction and the Guardian-executed destination-chain transaction.
Yes — in April 2025 a single Guardian went offline, which delayed Bitcoin deposit-address generation and withdrawals platform-wide. It wasn't a fund-loss exploit, but it showed that with only three Guardians, any single node's downtime can stall the entire pipeline.
Sources: Unit (HyperUnit), Architecture: Security — HyperUnit Docs, Deposited via Monad network — Hyperliquid FAQ, Hyper Unit — the hidden engine powering Hyperliquid — OAK Research, How do fees on Unit work? — Unit Docs
Last reviewed 2026-07-08. More Monad research.
Monad DeFi board · Markets · Swap on NullTerminal